Polarized P-glycoprotein expression by the immortalised human brain endothelial cell line, hCMEC/D3, restricts apical-to-basolateral permeability to rhodamine 123

Abstract

P-glycoprotein (P-gp) expression at the blood–brain barrier prevents unwanted blood-borne toxins and signalling molecules from entering the brain. Primary and immortalised human brain endothelial cells (BECs) represent two suitable options for studying P-gp function in vitro. The limited supply of primary human BECs and their instability over passage number make this choice unattractive for medium/high throughput studies. The aim of this study was to further characterise the expression of P-gp by an immortalised human BEC line, hCMEC/D3, in order to evaluate their use as an in vitro human blood–brain barrier model. P-gp expression was stable over a high passage number (up to passage 38) and was polarised on the apical plasma membrane, consistent with human BECs in vivo. In addition, hCMEC/D3 cell P-gp expression was comparable, albeit slightly lower to that observed in primary isolated human BECs although P-gp function was similar in both cell lines. The P-gp inhibitors tariquidar and vinblastine prevented the efflux of rhodamine 123 (rh123) from hCMEC/D3 cells, indicative of functional P-gp expression. hCMEC/D3 cells also displayed polarised P-gp transport, since both tariquidar and vinblasine selectively increased the apical-to-basolateral permeability of hCMEC/D3 cells to rh123. The results presented here demonstrate that hCMEC/D3 cells are a suitable model to investigate substrate specificity of P-gp in BECs of human origin.

Introduction

ATP-binding cassette (ABC) transporters at the blood–brain barrier (BBB) prevent the entry of unwanted endogenous signalling molecules and toxic xenobiotics from the blood into the brain [reviewed in Loscher and Potschka (2005)]. To date, at least 48 ABC transporter genes have been identified and are classified into seven subfamilies, denoted ABCA to ABCG (Dean et al., 2001, Higgins, 1992). ABCB1/MDR1 [(P-glycoprotein (P-gp)] is the most extensively studied ABC transporter at the BBB, both in vitro and in vivo, and prevents a broad range of substances from entering brain tissue, including many therapeutic drugs in use today such as anti-epileptics, anti-cancer drugs and HIV protease inhibitors [reviewed in Loscher and Potschka, 2005, Schinkel, 1999, Zhou, 2008].

Structurally, P-gp is arranged into two halves connected by a 75 amino acid linker region, with each half consisting of a transmembrane domain (made of 6 transmembrane segments), an intracellular nucleotide-binding domain and both the N and carboxyl termini, which are located intracellularly (Linton, 2007, Schinkel, 1999, Schinkel and Jonker, 2003). Conflicting evidence exist as to the expression of P-gp by brain endothelial cells (BECs). P-gp expression has been described on the luminal membrane of rat brain capillaries (Beaulieu et al., 1997, Miller et al., 2000), isolated mouse brain endothelial cells (Tatsuta et al., 1992), and human brain microvessels (Seetharaman et al., 1998, Virgintino et al., 2002). In contrast, enhanced abluminal expression of P-gp has been observed in rat BECs using transmission electron microscopy (EM) techniques (Bendayan et al., 2006). Nevertheless, it is widely accepted that P-gp expression is primarily located on the luminal membrane of BECs.

Whilst multiple studies have been carried out to identify the characteristics of P-gp substrates, the most common feature is their hydrophobic nature, which allows them to diffuse across the lipid bilayer relatively easily (Cabrera et al., 2006, Cianchetta et al., 2005, Crivori et al., 2006, Penzotti et al., 2002, Wang et al., 2003). Indeed, P-gp has been hypothesised to act as a flippase transporter (Schinkel, 1999). In vivo data supports the notion of a barrier function for P-gp at the BBB. Mice, lacking either mdr1a or combined mdr1a/b genes (equivalent to human ABCB1), display enhanced brain accumulation of many P-gp substrates including verapamil, indinavir, vinblastine and morphine, when compared to wild type control mice [reviewed in Loscher and Potschka (2005); Miller et al. (2008); Schinkel (1999)]. In addition, P-gp inhibition in rats results in increased brain concentrations of colchicine and vinblastine (Drion et al., 1996).

In vivo studies represent an important data source for evaluation of ABC transporter function at the BBB, but they are highly costly and extrapolation from animal models to humans may be problematic due to species differences. For routine investigations into P-gp function by human BECs in vitro, there are currently two choices, primary and immortalised human BECs. Human primary cell cultures are not readily available and, in our experience, do not remain stable over ∼ 5 passages, and are therefore not suited for routine medium/high throughput studies. hCMEC/D3 cells represent the most extensively characterized immortalised human BEC line to date, and retain a BEC typical phenotype (Weksler et al. 2005). hCMEC/D3 cells express tight junctional proteins and are restrictive to molecules larger than 4 kDa in molecular weight, in agreement with in vivo studies (Weksler et al. 2005). Previously, P-gp expression by hCMEC/D3 cells has been demonstrated using western blotting and mRNA analysis (Weksler et al. 2005). In the same study, an increase in the net influx of rhodamine 123 (rh123) was reported following inhibition of P-gp with PSC-833, demonstrating that the transporter was indeed functional, as confirmed in a recent study (Poller et al., 2008).

The aim of this study was to further examine P-gp expression and function in hCMEC/D3 cells to determine their suitability as an in vitro human BBB model for identifying potential P-gp substrates. We first investigated P-gp expression by hCMEC/D3 cells, including its subcellular localisation and polarisation, and compared it to that of primary human BECs. P-gp functionality in hCMEC/D3 cells and primary human BECs was then investigated using rhodamine 123 (rh123) efflux, with the 3rd generation P-gp inhibitor tariquidar, or the P-gp competitive substrate vinblastine. Finally, using P-gp inhibitors, we investigated the permeability of hCMEC/D3 cells to rh123 in order to confirm polarised P-gp function.